LongCat-Video
What it is
LongCat-Video is a 13.6B parameter foundational video generation model released by Meituan. It unifies text-to-video, image-to-video, and video-continuation tasks within a single architecture, utilizing a coarse-to-fine generation strategy and Block Sparse Attention to enable efficient, high-quality long-form video synthesis.
The problem it solves
Existing open-source video models often struggle with temporal consistency over long durations, suffer from quality degradation in extended sequences, or lack a unified architecture for multiple generation modalities.
What you could use it for
- Generating minutes-long coherent videos from text prompts
- Animating static images into dynamic video clips
- Extending existing video clips with consistent style and content
- Audio-driven human video generation via the LongCat-Video-Avatar variant
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.